The primary goal of this project is to provide the infrastructure for the Critical Assessment of Protein Structure Prediction (CASP) prediction experiments. With success of the sequencing projects, the structural characterization of the genome-encoded proteins becomes increasingly important. Extensive knowledge of protein structure will significantly aid the investigation of protein function, protein interactions, and biochemical pathways. It will also have a major impact on our understanding of biology, human disease, and eventually on drug design. Experimental determination of structure is inherently time-consuming and costly, and currently there is a two order-of-magnitude gap in numbers between sequencing and structure determination efforts. Computational techniques of structure modeling and prediction hold great promise for narrowing this gap. The CASP process was established to answer two questions: First, what level of prediction quality can be expected of these techniques? And second, which methods offer the most promise for continued development? CASP is a community-wide program, with 187research groups world-wide submitting over 28,000 predictions in the last round. Our group is the primary infrastructure resource for CASP, and handles processing of predictions, develops and implements evaluation software, performs prediction assessment, develops analysis and display tools, and facilitates access to predictions and their evaluation data. We propose to support the continuing operation of CASP and to expand its infrastructure, including an increased capacity for assessing predictions, further development and refinement of the evaluation software, and improved prediction analysis methods. We will implement a continuous web-based mechanism for evaluation of prediction methods, based on structural genomics prediction targets. We will also continue to provide the infrastructure for the Ten Most Wanted community-wide program to build useful structural models of biologically important proteins.